With the development of the Internet, it has become quite convenient for users to exchange messages over the Internet. Users can exchange various forms of messages through memo, chatting, messenger, and mail functions, etc., and the exchange of messages has become an essential feature in community websites and online games.
Although message transfer services are provided for users to exchange ideas, the number of occurrences is increasing in which these services are abused, at the cost of the users or the service providers.
For example, such cases may include obscene messages transferred to random users, and transaction messages sent to users to exchange game money obtained through hacking for cash.
In prior art, such spam messages were generally filtered using string matching methods. For instance, if ‘money’ was selected as a prohibited word to block the exchange of game money through messages, those messages which include the text ‘money’ were blocked. String matching is an algorithm based on artificial intelligence, and various string matching methods have been disclosed to date.
However, the number of spam message transmissions has increased, in which special characters such as symbols are used to elude string matching message filtering.
FIG. 1 illustrates an example of a message that uses symbols to elude string matching, and as in FIG. 1, symbols such as ‘{circle around (m)}’ and ‘{circle around (n)}’ are used to edit a letter sequence that can be discerned as the prohibited word ‘money’, whereby the number of occurrences is increasing in which messages that include prohibited words are transmitted.
These special characters can be used in combination with regular letters to form various letter sequences, and besides symbols, foreign letters known to the public can also be used to form words corresponding to the prohibited words. Thus, it is impossible to perform filtering for these words in actuality using the conventional string matching method.